A collection of software that I have created, or helped to create:
- Crossview Semantic
Transfer: Tensorflow code
for our CVPR 2017 paper, Predicting Ground-Level Scene Layout from
(Menghua Zhai, Zachary
- Deep Horizon Line
model and network definitions for estimating the horizon line from
a single image. This code accompanies our paper: Horizon Lines in
the Wild. (Scott Workman, Menghua Zhai,
Nathan Jacobs), In British Machine Vision Conference (BMVC), 2016.
- Deep Transient Attribute
Estimation: Caffe model
and network definitions for estimating transient attributes from a
single image. This code accompanies our paper: A Fast Method for
Estimating Transient Scene
Baltenberger, Menghua Zhai, Connor Greenwell, Scott Workman,
Nathan Jacobs), In IEEE Winter Conference on Applications of
Computer Vision (WACV), 2016.
- Deep Crossview
example code for our ICCV 2015 paper: Wide-Area Image
Geolocalization with Aerial Reference
Workman, Richard Souvenir, Nathan Jacobs), In IEEE International
Conference on Computer Vision (ICCV), 2015.
- Detecting Vanishing Points using Global Image Context in a
code for our CVPR 2016
by the lead author, Menghua “Ted” Zhai.
Some of these, and a few other smaller projects, are accessible from
You are free to use this software for educational purposes or academic
research. If you use any of this software please make sure to
acknowledge the relevant work. If you intend to use this in a
commercial application please contact me.